On Mar 31 18:48, Jon TURNEY wrote:
> On 30/03/2015 19:47, Corinna Vinschen wrote:
> >Just for the records what we talked about on IRC:
> >
> >On Mar 30 18:32, Jon TURNEY wrote:
> >>@@ -45,7 +49,7 @@ struct _fpstate
> >>    __uint32_t padding[24];
> >>  };
> >>
> >>-struct ucontext
> >>+struct mcontext
> >
> >__mcontext so as not to pollute the namespace.
> >
> >>    __uint64_t etr;
> >>    __uint64_t efr;
> >>    __uint8_t _internal;
> >>-  __uint64_t oldmask;
> >>  };
> >
> >Remove _internal, keep oldmask.  As a result, __mcontext is still
> >basically equivalent to Linux' mcontext_t.  __mcontext can be
> >taken from _my_tls.oldmask.
> 
> Thanks for your help with this.
> 
> You'll have to help me understand what the difference in meaning between
> ucontext_t.uc_sigmask and ucontext_t.uc_mcontext.oldmask is.
> 
> In the context of _cygtls::call_signal_handler() is _my_tls.oldmask correct
> and not this_oldmask?

Yes, this_oldmask should be the right one.


Thanks,
Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Maintainer                 cygwin AT cygwin DOT com
Red Hat